Viruses in the Mammalian Male Genital Tract and Their Effects on the Reproductive System

Transcription

1 MICROBIOLOGY AND MOLECULAR BIOLOGY REVIEWS, June 2001, p Vol. 65, No /01/$ DOI: /MMBR Copyright 2001, American Society for Microbiology. All Rights Reserved. Viruses in the Mammalian Male Genital Tract and Their Effects on the Reproductive System NATHALIE DEJUCQ* AND BERNARD JÉGOU GERM-INSERM U435, Université de Rennes I, Rennes, France INTRODUCTION VIRUSES AND SEMEN Human Semen Human immunodeficiency virus type Human T-lymphotropic virus type I Human herpesvirus Cytomegalovirus Epstein-Barr virus Papillomavirus Hepatitis viruses Herpesvirus and adenovirus Animal Semen VIRUSES AND THE TESTIS The Human Testis Mumps virus Human immunodeficiency virus type Oncogenic viruses Other viruses Endogenous retroviruses The Animal Testis Viruses and the seminiferous epithelium Viruses and the interstitial compartment VIRUSES AND THE PROSTATE AND OTHER ACCESSORY GLANDS Human Prostate Papillomavirus Herpes simplex virus Cytomegalovirus Human immunodeficiency virus type Human herpesvirus Human Seminal Vesicles and Epididymis Animal Prostate Animal Seminal Vesicles and Epididymis TESTICULAR ANTIVIRAL DEFENSE SYSTEM CONCLUSION ACKNOWLEDGMENTS REFERENCES INTRODUCTION Concern about the sexual transmission of viruses in humans and its health consequences has peaked with the appearance and development of the AIDS pandemic. There is also much concern about the possibility of sexual transmission of viruses in animals, particularly in economically important farm animals. Such sexually transmissible diseases (STDs) may cause epidemics, which may be spread further by the worldwide export of the gametes and embryos of these animals. In light of these major health and economic issues concerning sexually transmitted viral diseases (Table 1), it is paradoxical that, to * Corresponding author. Mailing address: GERM-INSERM U435, Campus de Beaulieu, Rennes Cedex, France. Phone: Fax: nathalie.dejucq@rennes.inserm.fr. our knowledge, no general review has been published concerning the presence of viruses in human and animal reproductive tracts and semen and the possible effects of these viruses within infected organs and on reproductive endocrinology. This review aims to address this subject by describing the various viruses identified in the semen and reproductive tracts of mammals, their distribution in tissues and fluids, their possible targets, and the functional consequences of their infectivity on the reproductive and endocrine systems. VIRUSES AND SEMEN Human Semen The viruses present in human semen and their consequences are listed in Table 2 and Fig

2 VOL. 65, 2001 VIRUSES AND THE MALE REPRODUCTIVE SYSTEM 209 TABLE 1. Matters of concern relating to viral infection of the reproductive tract or semen Spreading of diseases Infertility/sterility resulting from: Changes in one or more testicular compartments (e.g., germ cells, Sertoli cells, Leydig cells) Infiltration into the reproductive tract or semen of leukocytes causing a T-cell-mediated response to spermatozoa Cachexia induced by a drop in testosterone production Incorporation of the viral genome into the germ cell genome (risk of transmission to subsequent generations) Infection of ova and embryo, miscarriage, and embryonic and fetal abnormalities Human immunodeficiency virus type 1. Human immunodeficiency virus (HIV) is now the most extensively studied sexually transmitted virus. Its presence in the semen was rapidly established, but the nature of the cells infected remains unclear. One key question is whether the virus strains present in semen arise from the same compartment as those detected in blood cells, since new multidrug therapies are still unable to eradicate the virus completely (71). The virus was initially isolated from the mononuclear cell fraction of the semen of two men in the process of developing the disease and from one HIV-1-seropositive man (142, 355, 356). It has been shown to be transmitted via the semen of asymptomatic carriers (14, 19, 311). HIV-1 has also been detected in the cell-free seminal fluid of both an AIDS patient and an asymptomatic HIV-positive individual (47). Detection of the virus in this location led to the suggestion that the epididymal epithelial cells may become infected and release HIV into the epididymal fluid. The prostate and seminal vesicles may also act as a virus reservoir and release HIV into the semen. This contamination would occur in addition to that of macrophages and lymphocytes, which are often present in the semen and are natural targets for HIV (9). The possible presence of HIV-1 in the spermatozoa themselves is now a matter of debate and has recently been reviewed (8), raising questions about the vertical transmission of HIV. This problem is very acute because spermatozoa from HIV-positive men, cleared of seminal plasma and infected mononuclear cells, are used for TABLE 2. Presence of viruses in human semen and its consequences medically assisted reproduction in serodiscordant couples ( ). Thus, although semen washing on a density gradient before artificial insemination may reduce the risk of HIV transmission from the infected man to an uninfected woman (166, 188), the virus may still be detected in the fraction of motile spermatozoa used for insemination (70, 203, 316). Insemination in such cases is, of course, not performed. These positive results may be false positives due to the use of a single set of primers in the PCR detection of HIV (203) or due to the presence of contaminating cells of nonseminal origin (the mean proportion of such cells is 1/1,000) (316). However, electron microscopy and immunocytochemistry studies have provided evidence that HIV-1 can attach to the surface of spermatozoa and enter these cells through the intact plasma membrane (24 26). An in vitro study has also demonstrated that spermatozoa from healthy donors may carry HIV-1 on their surface and subsequently transmit it to lymphocytes in culture (104). Although still controversial (258), several lines of evidence suggest that HIV is able to bind to spermatozoa. However, it is unclear whether the virus penetrates and replicates in these cells or simply integrates into them. In another study (23), in situ PCR was used to determine HIV-1 provirus levels in seminal cells from 94 HIV-1-infected men at various stages of the clinical disease. Both seminal mononuclear cells and spermatozoa (35 and 33% of samples studied, respectively) were found to harbor HIV-1 proviral sequences. HIV DNA was also detected by PCR in the motile sperm fraction of all three samples tested (288). However, these results are not consistent with those of three other studies using PCR to detect HIV DNA (22, 213, 259). Experimental HIV penetration into the spermatozoa of healthy donors has been reported by Baccetti et al. (22), who used immunocytochemistry and in situ hybridization with electron microscopy to detect antigens and viral RNA. HIV RNA was detected within spermatozoa, but no viral DNA was detected in this study, indicating that if HIV did penetrate spermatozoa, it did not integrate and replicate within these cells. Consistent with this, it has been found that spermatozoa do not represent a significant source of HIV in semen (353). Indeed, a study has shown that vasectomy has little effect on the infectivity of semen, leading to the conclusion that most cell-free HIV in seminal plasma arises distal to Virus-infected cells and semen fractions (reference) Monocytes/macrophages and lymphocytes HIV (9, 142, 356) CMV (265) HBV (84, 135) HTLV-1? (318) Spermatotozoa HIV? (22, 26, 104, 203) HBV (84) HSV (172) Cellular fraction (no specific cell type identified) Papillomavirus (181, 183, 184) Adenovirus (80) Abnormalities detected in the presence of virus Infertility HSV (109, 172) Adenovirus (80) Azoospermia, oligospermia HIV (205) HSV (172) Morphologically abnormal spermatozoa HIV (AIDS) (175) Hematospermia CMV (169) Pyospermia HIV (AIDS) (175) Decrease in the number of CD4 cells CMV (190) Asthenospermia Papillomavirus (183)

3 210 DEJUCQ AND JÉGOU MICROBIOL. MOL. BIOL. REV. FIG. 1. Summary of the viruses found in the genital tract and semen of mammals.

4 VOL. 65, 2001 VIRUSES AND THE MALE REPRODUCTIVE SYSTEM 211 the vas deferens (176). Recently, semen samples from 52 men, 21 of whom were receiving antiviral therapy, were tested for HIV and the amount of virus present was quantified (316). HIV RNA was detected in 86% of semen plasma samples and in 14% of spermatozoon fractions (as stated above, the authors suggested possible contamination by nonspermatozoon cells at a mean frequency of 1/1,000), whereas HIV DNA was present in only 57% of nonspermatozoon cell fractions. It is unknown what prevents HIV replication in spermatozoa and which receptors are used by the virus to enter these cells. The CD4 receptor is present on semen lymphocytes and monocytes (125, 126), but it is unclear whether it is present on spermatozoa. Ashida and Scofield (21) were the first to describe a sperm ligand that reacts with CD4 antibodies and interactions between sperm and HLA-DR-positive cells, providing strong evidence that CD4 is expressed on spermatozoa. CD4 molecules were detected on mouse sperm heads by both immunofluorescence and western Blotting (189). Their presence was also suggested, indirectly, in a study using a semen protein (gp17) that binds to CD4 T cells and soluble recombinant CD4, as well as spermatozoa (38). However, the CD4 antigen has never been detected on the surface of human spermatozoa or on CD45 ejaculate cells (epithelial and germinal cells) (15, 110, 125, 166, 240, 352). Neither of the two main HIV coreceptors (CXCR4 and CCR5) were detected on the surface of spermatozoa in flow cytometry experiments (166), although the possibility that they are expressed at a very low level cannot be ruled out. It has been suggested that other receptors are responsible for the entry of HIV-1 into spermatozoa. Thus, several studies have described a glycolipid that may function as an HIV receptor on the surface of spermatozoa and that may be involved in transmission of the virus (21, 22, 52, 53, 289). Sperm proteins that bind HLA-DR and are therefore likely to interact with somatic cells have also been described (276, 289). HIV may also infect germ cells early in spermatogenesis, resuting in the clonal transmission of the virus into spermatozoa; this is discussed below (see Viruses and the testis, below). The effect of HIV infection on semen characteristics has been investigated. HIV was isolated from 15 (30%) of 50 specimens from asymptomatic individuals and from 1 of 3 specimens from patients with AIDS (175). The men with AIDS all had pyospermia and grossly abnormal spermatozoa. In contrast, the semen specimens from other seropositive men did not differ significantly from those of healthy seronegative donors. No abnormality in sperm count, morphology, number or type of leukocytes in semen, or any other semen characteristic was associated with HIV shedding into semen. In another study, a significant positive correlation was found between blood CD4 cell number and sperm motility in seropositive men, and a significant inverse correlation was found between CD4 cell number and sperm abnormalities. The authors suggested that this may be due to a decrease in testosteronemia, resulting in defective epididymal sperm maturation (102). Several studies have investigated whether the level of HIV-1 in semen varies with the stage of infection. It has been reported that the isolation of HIV from semen does not correlate with CD4 or CD8 T-lymphocyte counts and that seropositive men may shed HIV in semen early in the course of infection (175). Other studies have also reported that the presence of HIV DNA in semen is not related to the CD4 cell count or disease status (198, 213). Indeed, although HIV-1 is more common in the semen of men with advanced HIV-1 infection and seminal leukocytosis, it can also be isolated from the semen of men with neither of these conditions (16, 342). Furthermore, men with HIV-1 infection are already potentially infectious through sexual relations during the first few weeks after infection (329). It has now been established that HIV-1 may be present in semen in both cell-free and cell-associated forms (and that it may be isolated from both asymptomatic individuals and AIDS patients) and that both forms are transmissible (360). Surprisingly, HIV seems to be shed intermittently into semen (50, 174). Concomitant STDs such as cytomegalovirus (CMV) (174), chancroid, syphilis, gonorrhea, and Chlamydia infections (112) may affect the level of HIV shedding (206, 312). Herpes simplex virus (HSV) increases plasma HIV levels severalfold (221), and this increase may be reflected in seminal fluid. Also, the membrane proteins of CMV and human T-lymphotropic virus type I (HTLV-I) have large regions of similarity to CD4 (276), suggesting that cells infected by these viruses may be more susceptible to HIV infection. Cohen et al. (72) compared two groups of HIV patients, consulting for dermatological problems or genital infections (urethritis). The median HIV RNA level in the semen of the group with dermatological infections was one-eighth that in the group with genital infections. In the group with genital infections, the subgroup of patients with gonorrhea had the highest seminal viral load. Antibiotic treatment of urethritis reduced the viral load in the semen but did not affect the plasma viral load. This study clearly establishes that local infections of the male reproductive tract are important cofactors of HIV load in the semen. The question of viral compartmentalization was raised in a longitudinal analysis of eight subjects who went on to develop AIDS. The seminal viral load increased in most cases, but the viral load was consistently higher in blood plasma than in semen (132). An absence of correlation between plasma and semen loads was also reported in another study and suggests that the semen and blood are separate viral compartments (198). A comparison of HIV-1 gp120 sequences from five recent seroconverters with those from their corresponding sexual partners (transmitters) revealed that in each couple studied, the variant transmitted corresponded to a minor population in the semen of the transmitter, providing evidence that HIV-1 selection occurs during sexual transmission (360). Protease gene sequences also differ in semen and blood (55, 165), as do viral phenotypes (342) and the ratio of infected to uninfected leukocytes (165). Further evidence of viral compartmentalization is provided by the following observations: the lack of association between the culturability of the virus in semen and viral RNA levels in blood, the discordant distributions of viral phenotypes, the discordant viral RNA levels, the absence of correlation between viral RNA levels in semen and CD4 cell counts in blood, differences in the biological variability of viral RNA levels, and differences in viral load following antiretroviral treatment (75). HIV-1 comprises two main phenotypic strains: NSI (non-syncitium inducing) strains, which infect macrophages, use CCR5 as a coreceptor for cell entry, and are preferentially transmitted; and SI (syncitium-inducing) strains, which use the CXCR4 coreceptor, appear later in the disease,

5 212 DEJUCQ AND JÉGOU MICROBIOL. MOL. BIOL. REV. and are poorly transmitted. It has therefore been suggested that there is selection between NSI and SI strains in the genital tract. However, restriction of SI variants in the male genital tract, such as would account for the observed NSI transmission bias, remains to be established, because SI and NSI strains do not seem to be compartmentalized in semen (90). The precise identification of the viral reservoir in the body is now of the utmost importance for antiretroviral treatment. Thus, although the development of potent treatments raises the hope that HIV-1 eradication might be possible, we do not yet know whether all the compartments in which the virus replicates are accessible to antiretroviral compounds. Several studies have investigated whether drug-resistant strains develop in seminal plasma and whether patients undergoing treatment carry infectious variants in their semen. A study by van t Wout et al. (340) analyzed the relationship between HIV-1 quasispecies extracted from the blood at various times and from various organs at autopsy. The brain quasispecies were very homogeneous and differed from the peripheral blood mononuclear cell variants, suggesting compartmentalization and the early spread of HIV-1 to the brain. Tissue-specific quasispecies were also observed in the testis, but in this case they were suggestive of a later invasion, possibly secondary to lymphocyte infiltration due to the disease (340). An evaluation of blood and genital secretions from HIV-infected men under treatment showed no genotypic changes consistent with protease inhibitor resistance in semen, despite the presence of these agents in blood plasma. This therefore suggests that protease inhibition may have limited penetration into the male genital tract (111). Replicationcompetent virus was detected in the semen of HIV-infected men receiving antiretroviral treatment, although there was no detectable virus in the peripheral plasma (358). In light of these studies, it appears that treatment strategies for the complete elimination of HIV-1 from the genital tract should now be a public health priority. It is therefore urgent to study the penetration of antiretroviral compounds into the male genital tract and to establish a correlation with the presence of virus and virus subtypes. Human T-lymphotropic virus type I. HTLV-I, like HIV, is a retrovirus that infects T cells. It is epidemiologically linked to adult T-cell leukemia-lymphoma (230). HTLV-I is sexually transmitted by semen (231, 317, 318), most probably via contaminated lymphocytes in the semen. Human herpesvirus 8. Kaposi s sarcoma (KS) is frequently associated with AIDS and occurs mainly in homosexual men. The epidemiology of KS in HIV-infected patients suggests that it may be caused by a sexually transmitted infectious agent (37, 144). This agent has recently been identified as a new human herpesvirus called human herpesvirus 8 (HHV-8) or KS-associated herpesvirus. According to many studies using extremely sensitive serologic techniques (76, 97, 133, 145, 223, 321, 322), the prevalence of HHV-8 in the general population is low. However, it is generally accepted that the virus can be detected in the semen of KS-positive patients. A recent study detected HHV-8 DNA in 12% of semen samples from KS patients, with a mean copy number per microgram of positive target DNA of 300, versus 9,000 in blood cells (182). Significant amounts of HHV-8 DNA were also detected in semen samples from 28 HIV-1-infected individuals and were shown to infect the mononuclear cell fraction (45). A recent multicenter comparison study concluded that HHV-8 DNA is present in semen but at concentrations that are probably too low to facilitate its consistent detection (251). Thus, the semen viral load should be measured to determine whether it is high enough for sexual transmission to occur. In addition, the prevalence of HHV-8 in the semen of healthy men has not yet been accurately established. It should, however, been borne in mind that the incidence of KS in HIV-negative individuals is much higher in Italy than, for example, in the United Kingdom (30 times higher) or the United States (20 times higher). In Denmark, a recent study performed on 100 healthy donors did not detect any seminal HHV-8 DNA (161). Therefore, the prevalence of HHV-8 in semen may be higher in some geographical areas than in others. The nature of the infected cells in semen and the origin of the virus also remain to be determined. Cytomegalovirus. CMV also belongs to the Herpesviridae family. It is very common, with 50% of the otherwise healthy population being infected (349). Infection is spread by intimate contact with infected body fluids including semen (185), and 40% of the semen from healthy donors is infected. This virus was previously erroneously reported to be cold labile, whereas it in fact survives in frozen and thawed semen (137). CMV is thought to be a possible causative agent of hematospermia (169). The virus generally remains in a latent form and causes a lifelong infection, but it may be activated either by a primary infection, for example after organ transplantation, or by the impairment of cellular immunity. Prospective studies in the United States have demonstrated that CMV is responsible for more prenatal and perinatal virus infections than is any other transmissible agent identified to date. The impact of these infections on fetal and neonatal health is unclear. However, preliminary data suggest that CMV is probably the most important agent responsible for congenital infection and damage (185). The French government decided some time ago to reject CMV-seropositive donors for artificial insemination. It then pulled back from this decision, deciding instead to reject only donors with recent seroconversion. Two independent laboratories tested for CMV in 178 cryopreserved sperm samples from 97 healthy donors, 34% of whom were CMV seronegative, 52.6% of whom were seropositive with no recent contamination (absence of immunoglobulin M), and 13.4% of whom had unknown serological status. They detected CMV in 2.8% of the samples after culture and in 5.6% by PCR, thereby demonstrating that CMV can be detected in the semen in the absence of recent contamination (200). CMV is also one of the most common opportunistic infections in AIDS patients, and it has been suggested that persistent CMV infection of the semen increases the risk of AIDS, possibly by activating CD4 cells such that HIV-1 is produced (93). CMV has been isolated from the semen of homosexual men (202, 301), into which it was excreted intermittently (73), and CMV levels are associated with HIV seropositivity (186, 270). However, Rinaldo et al. found no association between CMV shedding and a higher risk of developing AIDS (269). The shedding of HIV was more closely associated with the concomitant shedding of CMV than with the CD4 cell count (174). In a study investigating the relationship between CMV infection and the progression of HIV-1 disease, a group of 234 asymptomatic HIV-1 antibody-positive homosexual men were

6 VOL. 65, 2001 VIRUSES AND THE MALE REPRODUCTIVE SYSTEM 213 tested for CMV. CMV was isolated from the semen of 45% of the men. CD4 cell levels were significantly lower in those in whom CMV had been isolated from semen. Similarly, an inverse relationship was observed between the concentration of CMV in semen and the CD4 cell levels (190). Leach et al. (191) later concluded from junctional hybridization experiments that the presence of multiple CMV strains in HIV-1- positive homosexual men was associated with the progression to AIDS, possibly via activation of HIV-1-infected CD4 cells. Hematopoietic cells are the only cells in the semen to have been identified as being infected (265). Epstein-Barr virus. Epidemiological studies have shown that Epstein-Barr virus (EBV) infection is most common in the age group at which sexual activity begins, strongly indicating that it is sexually transmitted (334). The presence of EBV in semen has not yet been investigated, but several studies have reported that human seminal plasma activates replication of this virus. Thus, EBV early-antigen expression is stronger in infected cells cultured in the presence of semen (149, 151, 330, 357). In vitro inhibition by seminal plasma of both the T- and B-lymphocyte responses to infection has also been described (193, 334). These results suggest that seminal plasma may facilitate EBV replication in the cervix of the uterus and may therefore have some relevance to the etiology of cervical cancer. Papillomavirus. Papillomaviruses, like EBV, cause cancer. There has been extensive testing for these viruses in the male genital tract, especially in the prostate, since they are thought to be a possible cause of prostate cancer; this is discussed below (see Viruses and the prostate and other accessory glands ). Human papillomavirus (HPV) DNA was detected in the semen of three patients, and the data obtained were consistent with the contention that HPV can be transmitted sexually via semen, as suggested by epidemiological data on the sexual transmission of HPV (247). Since that first report, papillomaviruses have been detected in semen in several studies (130, 131, 150, 181, 184). These results conflict with early work by Nieminen, which claimed that papillomavirus DNA was not transmitted by semen since the semen of the sexual partners of 17 women positive for HPV DNA was uninfected (238). However, these results actually show only that transmission from the woman to the man is rather inefficient. Indeed, other studies have indicated that HPV type 16 and 18 DNA is present in sperm cells and may be transmitted to the partner (64, 181, 183, 184). Thus, HPV type 16 DNA and RNA were detected in the semen of 25 and 8% of 24 randomly selected patients, respectively, whereas the prevalence of detection for HPV type 18 DNA and RNA was higher (46 and 21%, respectively). The incidence of asthenozoospermia is significantly higher in patients with HPV in their semen (183). Hepatitis viruses. Hepatitis viruses may cause acute diseases (fulminant hepatitis) or chronic diseases such as liver cancer and cirrhosis. The ability of human semen to transmit hepatitis B virus (HBV) was first demonstrated by the inoculation of gibbons (291). HBV antigens were subsequently detected in human semen (155), and it is now well established that this biological fluid is a vector for the spread of hepatitis B (84, 107, 115, 153, 158). However, few studies have tried to identify the contaminated cells within semen. Hadchouel et al. (135) showed that HBV DNA was integrated into the DNA of spermatozoa in two of three patients with acute hepatitis, suggesting that there may be true transmission of HBV via the germ line. Another study of chronic HBV antigen carriers showed that HBV DNA was present in all of the semen samples tested with the infected cells being both spermatozoa and mononuclear cells (85). Persistent free HBV DNA has also been detected in the semen of patients with no markers of viral replication in serum, indicating that the genital tract may act as a reservoir and that these patients may transmit the virus sexually (85). The transmission of hepatitis C virus (HCV) is a major health concern since the disease is asymptomatic in threequarters of cases. Half of the infected patients become chronic carriers, and 10% develop liver cancer or cirrhosis. Although HCV transmission via parenteral exposure is well documented, the sexual transmission of this virus is more contentious (for reviews, see references 275 and 347). To determine whether HCV could be sexually transmitted, the frequency of HCV infection was studied in heterosexuals with multiple partners. The frequency of HCV infection in these individuals was found to be much higher than that in healthy women with a stable partner (359). In Egypt, where the rate of seropositivity for HCV is particularly high (13 to 22% versus 0.04 to 1.2% in the United States and Europe), a study suggested that HCV is transmitted within couples (180). Thus, viral nucleotide sequence analysis for 33 husband-and-wife pairs revealed significantly high levels of similarity (97 to 100%) in 32 of the 33 pairs studied. Although the role of sexual contact in such transmission is unclear, these and other studies (5, 11 13, 114, 157, 196, 326) indicate that the sexual transmission of HCV cannot be ruled out. The search for HCV RNA in semen has generated conflicting results, with about half the studies demonstrating the presence of HCV in seminal fluid (173, 195, 197, 209, 319) and the other half demonstrating its absence (56, 121, 147, 296, 324) or finding its prevalence to be low (117). Thus, in the group of studies supporting the hypothesis of a sexual transmission, a study by Liou et al. showed by nested PCR that among 34 patients with chronic liver disease positive for anti-hcv antibodies and with HCV RNA in serum, the prevalence of HCV RNA in body fluids was 100% (7 of 7) in ascites, 48% (15 of 31) in saliva, 7% (2 of 29) in urine, and 24% (4 of 17) in semen (195). HCV-specific antigens have also been detected in semen (173), and HCV RNA was detected in the supernatant of spermatozoa and spermatids from five patients with chronic hepatitis C (197). In contrast, several studies argue against a sexual transmission of the virus, since they have detected no HCV RNA in the semen of chronically infected patients (56, 121, 147, 296, 324) or have found its prevalence to be low (117). The risk of hepatitis C transmission during artificial insemination has been investigated, and although HCV RNA was detected in the semen of some donors, no viral RNA remained following purification (209). In conclusion, although HCV may be present in semen, the viral load is probably extremely low, and therefore the virus represents a minor risk of sexual transmission. One group recently presented preliminary evidence that hepatitis G virus (HGV) is present in semen (296), whereas another detected no HGV RNA in semen samples obtained

7 214 DEJUCQ AND JÉGOU MICROBIOL. MOL. BIOL. REV. TABLE 3. Presence of viruses in animal semen and its consequences Virus and species infected (reference) Cells infected Effects BTV (bull) (51, 146, 250, 252) Spermatozoa (118) Semen abnormalities (118) early embryonic death, abortion, malformed fetal calves (245); transient infertility (245) BHV-1 (bull) (106) Seminal plasma rather than cells (337) Decrease in semen quality (106); fertility disturbances ( ) Foot-and-mouth disease virus (bull) (78, 292, 307) NA a Transmitted by artificial insemination (35, 58, 86, 285) BLV (bull) (271) NA Does not seem to be transmitted by semen; presence of a factor inhibiting virus replication? (35, 225, 271, 313, 332) BIV (bull) (234) Leukocyte fraction NA BDV (bull) (28) NA Sexual transmission (214, 215, 285); no apparent sperm defect (170) SIV (monkey) (218) Rarely recovered from mononuclear cells (217) Sexual transmission PRRSV (pig) (341) Macrophages (69) Sexual transmission (6, 237) Porcine parvovirus (pig) (128) DNA binding to spermatozoa (128) NA Porcine rubulavirus (pig) (263) NA Reduction in spermatozoon mobility and concentration (263) CMV (mouse) (31) Spermatozoa (31) Sexual transmission (235) Retroviruses (mouse) (162, 164, 254) Macrophages (164, 194) NA Spermatozoa (164, 194) FIV (cat) (154) NA NA (sexual transmission) Herpesvirus (gibbon) (36, 327) NA Sperm abnormalities Rous sarcoma virus (Xenopus laevis) (134) Spermatozoa Transfer to ova, developmental malformations (134) a NA, not available. from a cohort of 54 HIV-1-infected homosexual men (143). The pathological significance of HGV is unknown. Herpesvirus and adenovirus. HSV is a sexually transmitted agent that may be associated with cervical cancer and may cause high morbidity and mortality in perinatal infection (7, 233). Centifanto et al. (60) were the first to suggest that it was present in the male genital tract: inclusion bodies that resembled those of herpesvirus were found in tissue cultures of semen from two subjects, but no infectious virus could be cultured directly from the samples (94). Moore et al. (226) described an individual in a therapeutic donor insemination program who asymptomatically acquired a primary HSV-2 infection and transmitted it to one of two HSV-seronegative partners, in whom a primary HSV-2 infection developed, providing evidence for the sexual transmission of HSV-2. HSV DNA was recently detected in the semen of men with genital HSV-2 infection, mainly during recurrences of herpes (345). Concerning the type of cell infected within semen, an early electron microscopy study found no ultrastructural relationship between the virion and the spermatozoon (302). However, HSV DNA has recently been detected in human spermatozoa by in situ hybridization (172). The association of HSV DNA with spermatozoa has been linked to fertility problems, because HSV infection was found to be almost three times as common in the semen of patients with a low sperm count attending an infertility clinic as in individuals with a normal sperm count (172). A significant association was also found between infertility and positive tests for HSV in another study performed on 153 men (109). A possible relationship between infertility and the presence of the adenovirus in semen has also been a matter of concern. Csata and Kulcsar (80) detected HSV or adenovirus in the semen of 40% of infertile patients tested and found that these viruses were present in a latent form in 60% of their spermatozoa. Further studies are required to confirm these data and to eludidate the link between the presence of these viruses in semen and fertility problems. Animal Semen Several viruses have been detected in the semen of a number of animal species (Table 3). DNA from Rous sarcoma virus binds to Xenopus laevis spermatozoa and is transferred to ova during fertilization, inducing developmental malformations in 25 to 30% of embryos (134). In gibbons, sperm abnormalities are frequently encountered and have often been shown to be related to the presence of herpesvirus (36, 327). Feline immunodeficiency virus is shed into the semen of both experimentally and naturally infected cats (154). Simian immunodeficiency virus (SIV) is also present in the semen of monkeys and is transmitted by this vector (217). In mice, high concentrations of retroviral particles have been detected in the epididymal fluid ( ). These viruses are mostly endogenous retroviruses, but some exogenous infectious retroviruses (mouse ecotropic virus and Friend ecotropic virus) have also been described (254), and the association of retroviral particles with spermatozoa may lead to congenital infection. Murine leukemia virus has been found free in the seminal fluid, fixed to spermatozoa, and associated with macrophages (164, 194). The main site of virus synthesis within the male genital tract is the epithelial cells lining the epididymis duct. Murine CMV DNA has also been detected in spermatozoa, and the affected mice remained fertile (31). The porcine reproductive respiratory syndrome (PRRS) was

8 VOL. 65, 2001 VIRUSES AND THE MALE REPRODUCTIVE SYSTEM 215 first recognized in the United States in 1987 and in Europe in 1990 and has now spread worldwide. The PRRS virus (PRRSV) has been detected in semen (341) and is transmitted sexually (6, 237). Attempts have been made to determine the origin of PRRSV in semen by inoculating vasectomized and nonvasectomized boars intranasally with the virus. PRRSV RNA was detected in semen from all the boars and was most consistently found within macrophages. Therefore, macrophages in semen are probably infected via the infection of local tissue macrophages or via infected circulating monocytes or macrophages (69). In contrast, porcine parvovirus DNA, detected in the epididymal semen of oronasally inoculated boars, was found to bind to spermatozoa (128) but seemed to have no negative effect. Lastly, porcine rubulavirus has been reported to reduce sperm quality in sexually mature boars by decreasing spermatozoon motility and concentration (263). Since viral contamination of semen is common in bulls and since frozen semen is widely distributed and is of major economic importance, national and international organizations have laid down guidelines aimed at establishing disease-free bull studs producing semen free from potential pathogens. Many different viruses have been detected in bull semen. In particular, bovine diarrhea virus has been detected at high titers in semen (28) and is transmitted by semen (214, 215, 285). This virus does not seem to cause sperm defects (170). Bluetongue virus (BTV) has also been isolated from bull semen (51, 146, 250, 252), and a positive relationship has been found between the infectivity of semen samples from bulls infected with latent BTV and semen abnormalities (118), with virus-like particles occasionally observed in the heads of affected spermatozoa (118). BTV is associated with reproductive disorders including early embryo death, abortion, malformation of fetal calves and lambs, and transient infertility in bulls and rams coinciding with the shedding of virus into semen (245). Bulls were experimentally inoculated with BTV to investigate the frequency, duration, and pathogenesis of virus shedding into the semen. The shedding of BTV into semen in infected bulls closely followed peak virus titers in blood, and in no case was the virus isolated from semen without also being isolated from blood. Three of nine heifers inseminated with semen from a bull shedding BTV became viremic, seroconverted, and became pregnant. However, there was no evidence of fetal infection (48, 49). Bovine herpesvirus 1 (BHV-1), one of the most common viral pathogens in bovine semen, may also be associated with a decrease in semen quality, and bull semen is therefore systematically tested for this virus before being used for insemination (106). BHV-1 is present in the seminal plasma rather than the cell fraction (337). The risk of transmitting BHV-1 to cows by insemination has been investigated, and such transmission has been shown to lead to fertility disturbances ( ). Foot-and-mouth disease virus has also been detected in bull semen and is transmitted by artificial insemination (35, 58, 78, 86, 285, 292, 300, 307). In contrast, bovine leukemia virus is present in semen but does not seem to be transmitted by it (35, 225, 271, 313, 332), possibly due to the presence of an inhibitory factor in fresh semen that prevents bovine leukemia virus infection from becoming established (271). Finally, bovine immunodeficiency virus, a lentivirus, is associated with a lymphoproliferative disease and is prevalent in dairy and beef cattle in the southeastern United States. Its mode of transmission is unknown, but semen is suspected to be a potential vector of infection. Indeed bovine immunodeficiency virus DNA has been detected in the leukocyte fraction of cryopreserved semen specimens (234). VIRUSES AND THE TESTIS Semen is known to be an important vector in the dissemination of viral diseases, and several viruses are present in cell-free semen and in seminal cells including spermatozoa, macrophages and lymphocytes (see Viruses and semen above). Paradoxically, testing for viruses in the testis has not been extensive despite the possibility that they are involved in some of the disorders of this organ, such as orchitis, decrease in semen quality, azoospermia, and testicular carcinoma. The blood-testis barrier makes it likely that the testis acts as a reservoir for viruses (which may be protected against antiviral treatments), and this is another crucial reason for considering the testis to be an organ of special interest in the context of viral infection and STDs. Viruses present in the testis are listed in Fig. 1. The mammalian testis has two main compartments: (i) the interstitium, which contains the Leydig cells responsible for testosterone production, macrophages, fibroblasts, blood, and lymphatic vessels; and (ii) the seminiferous tubules, which are bordered by the peritubular myoid cells and are composed of the various generations of germ cells (from spermatogonia [the stem germ cells] to the meiotic spermatocytes and the haploid spermatids that differentiate into spermatozoa) that are associated with the somatic Sertoli cells (152). Viruses have been found in both these compartments in humans and other mammals. The Human Testis The best-known viruses causing testicular disorders in humans are mumps virus and HIV. A few studies have considered the role of oncogenic viruses, such as EBV or papillomavirus, in the etiology of testicular carcinoma even though testicular germ cell tumors are the most common malignant tumors in young adults and the incidence of testicular cancer has been steadily increasing in all countries for which epidemiological data are available (2, 39, 331). The viruses found in the human testis are listed in Table 4; see also Table 6. Mumps virus. Mumps is caused by an RNA virus of the paramyxovirus group. In prepubertal boys, the symptoms of mumps are usually limited to infectious parotitis, but, in men, orchitis is the most common complication (29, 34). Orchitis develops in 5 to 37% of all adult patients infected with mumps (120). Within the first few days of infection, the virus directly attacks the testes, destroying the testicular parenchyma (44, 242) and decreasing androgen production (4). This accounts for the testicular atrophy observed in 40 to 70% of patients with orchitis (29, 34). Unilateral involvement is the most common, while bilateral involvement occurs in 15 to 30% of the patients with orchitis (201). Bilateral orchitis leads to hypofertility with oligospermia and testicular atrophy in 13% of those patients (59). Morphological studies of mumps-associated orchitis were carried out in the 1940s by Gall (122) and Charny and Meranze (65). They showed the focused nature of the inflammation and described the sequential stages of the dis-

9 216 DEJUCQ AND JÉGOU MICROBIOL. MOL. BIOL. REV. TABLE 4. Viruses found in the human testis and their consequences Virus (reference) Cells infected Effects Mumps virus Leydig cells (4), germ cells? Orchitis (29, 34, 120), testicular atrophy (29, 34), sterility (3, 34, 159), decrease in androgen secretion (4), testicular cancer? (34, 243) HIV Lymphocytes and macrophages (257), germ cells (228, 240) Orchitis, interstitial fibrosis, lymphocyte infiltration, change in Leydig cell number, decrease in germ cell number (63, 220, 236, 266, 348), change in spermatogenesis (63, 82, 83, 92, 257, 273, 354) EBV (67) NA a Orchitis (268), testicular cancer? (10, 262, 305) Parvovirus B19 (129) NA Testicular cancer? (129) HSV-2 (95) NA Viral reservoir? (95) HSV-1 (80) NA Infertility? (80) Adenovirus (80) NA Infertility? (80) Coxsackie virus? NA Orchitis (216) Influenza virus? NA Orchitis (216) Endogenous retroviruses Germ cells (103, 187) Testicular cancer? (320) a NA, not available. ease. During the initial stage, an interstitial edema was commonly detected. Blood vessels were congested and surrounded by lymphocytes. Increased permeability of the blood vessels was found to lead to local interstitial hemorrhage and to exudation of leukocytes and fibrin. The seminiferous epithelium degenerated, but the Sertoli cells seemed little affected. Repair of the damage caused by infection involved the deposition of collagen within the interstitium, followed by tubular atrophy and peritubular concentric fibrosis, the usual residue of mumps orchitis. Sterility may be transient or definitive and is due to the loss of the germinal epithelium (290). There are several possible explanations for the mumps-induced degeneration of germ cells. Since mumps virus does not seem to induce germ cell transformation or proliferation in vitro, it may have an indirect effect (243): (i) high fever associated with the disease leads to a change in testicular temperature, contributing to germ cell degeneration (the most common hypothesis); (ii) germ cell degeneration may be caused by seminiferous tubule congestion following the interstitial edema (201); or (iii) a modification of testosterone production by Leydig cells may have a deleterious effect on seminiferous tubule function. Data concerning the consequences of orchitis on testicular endocrine function are rare. Adamopoulos et al. (3) described a severe alteration of Leydig cell function during the acute phase of the disease. These authors observed a drop in the testosterone level together with an increase in luteinizing hormone (LH) and follicle-stimulating hormone (FSH) levels in 27 patients suffering from mumps orchitis, suggesting a testicular failure. In another study (4) examining three patients, testicular atrophy, libido decrease, impotence, and gynecomasty were associated with the reduction of testosterone and the increase in LH and FSH levels. Thus, in some cases, the Leydig cell function seems to be damaged by the mumps infection. Whether this alteration is due to a direct or indirect effect of the virus on the cells is unknown. Such a dysfunction would also account for the higher incidence of testicular cancer in men who have had mumps than in control men (34). Two publications suggested that mumps virus replicates within the testis: Bigazzi et al. (42, 43) showed viral replication in the interstitial tissue in organotypic culture of monkey testis following virus inoculation through the testicular vein, and Bjorvatn et al. (44) rescued mumps virus by needle aspiration of testis biopsy specimens from mumps-infected patients. Two studies (113, 177) showed that systemic treatment with interferon (IFN) prevented the testicular atrophy caused by mumps orchitis in all treated patients (4 of 4 and 13 of 13, respectively). In the second study, where patients were randomly assigned into two groups, atrophy of the testes was observed in three of eight men in the control group. However, asthenospermia was still detected in four patients 7 years after interferon treatment (177). Human immunodeficiency virus type 1. Several endocrine and testicular dysfunctions have been reported in men infected with HIV-1, depending, in part, on the stage of the disease. High levels of testosterone have been found in men during the early stages of HIV infection (68, 212), whereas low testosterone levels have been found in men with AIDS (100, 261, 287, 343). Hypogonadism is common in HIV disease, and the decrease in testosterone levels certainly contributes to the weight loss observed in AIDS patients. Indeed, in men with low testosterone levels, testosterone replacement via a transdermal system is followed by a gain in lean body mass (41). The testosteronemia probably results from lymphocyte infiltration and fibrosis of the interstitial tissue and from a decrease in Leydig cell number (82, 83, 257, 273). AIDS patients often suffer from orchitis, hypogonadism, oligospermia, or azoospermia (20, 100, 253, 257, 325) and, in some cases, from testicular germ cell tumor or lymphoma (54, 116, 244, 351). Among 3,015 HIV-positive men, the incidence of testis tumors has been reported to be 0.2%, which is 57 times that of US. average of 3.5 cases/100,000 men (351). Several autopsy reports for deceased AIDS patients have shown a high rate of testicular abnormalities (63, 220, 236, 266, 348). Several groups have studied testis histology in men with AIDS and spermatogenesis dysfunction is invariably reported. This was first described by Chabon et al. (63), who showed major changes in the testes of 20 patients, resulting in a Sertoli cell only pattern, and interstitial inflammation in 53% of the patients. Rogers and Klatt (273) also described germ cell depletion, a decrease in tubule diameter, and an increase in the thickness of the tubule boundary wall. Yoshikawa et al. (354) set up a classification system for the disorders observed in the testes of 31 patients who had died from AIDS. Five categories were established: Sertoli cell only (42%), germ cell degeneration (27%), peritubular fibrosis associated with tubular hyalinization (15%), maturation arrest (12%), and normal appear-

10 VOL. 65, 2001 VIRUSES AND THE MALE REPRODUCTIVE SYSTEM 217 ance (3%). These observations were confirmed in 57 autopsy cases in another study (92). Since then, other authors have described the arrest of spermatogenesis at various points, numerous foci of degenerating germ cells, and epididymis block. Interestingly, a study of 140 testicular autopsy specimens from AIDS patients with and without antiviral treatment showed that treatment and prolongation of survival in AIDS patients is associated with a shift in the histologic findings for testes toward a more pronounced loss of germ cells (304). These morphological observations have raised questions about the mode of action of HIV on the testis. The virus was first thought to have an indirect effect due to the chronic debilitating illness and cachexia of the patients. It has also been suggested that opportunistic infections are involved. However, another study indicated that only 32% of patients with opportunistic infections had CMV, Mycobacterium avium-intracellulare, or Toxoplasma gondii in the testes. This suggested that these infections were probably not the key factors causing the observed hypogonadism (91). HIV may also act indirectly via changes in the hypothalamic-pituitary axis. Men with AIDS often have low testosteronemia (see above), and a dysfunction in hypothalamic gonadotropin releasing hormone secretion may account for this phenomenon (100). However, other studies found no significant abnormality in the hypothalamic-pituitary axis (253, 286). The low serum testosterone levels in men with AIDS are in some cases associated with high serum LH and FSH levels, implying that there is primary testicular failure (79). A possible reason for the testosterone secretory defect in these men is suggested by reports that cytokines released by the activated phagocytic cells of the immune system inhibit the steroidogenic response to human chorionic gonadotropin in vitro and presumably also the response to LH in vivo (57). Recent studies have shown that physiologic testosterone replacement in HIV-infected men with weight loss who have low testosterone levels can increase muscle mass and effort-dependent strength (for a review, see reference 40). Testosterone therapy also helps to alleviate the symptoms of hypogonadism (260). However, further studies are needed to determine whether androgen therapy can significantly and durably improve physical function and health-related outcomes in HIVinfected men. HIV infection of the testis has now been described in several studies, indicating that direct local action may be responsible for the observed damage within the gonads. The HIV p17 protein was first detected within the testis by immunohistochemistry using monoclonal antibodies (83). Pudney and Anderson (257) detected the CD4 receptor on the cell surface of lymphocytes and macrophages infiltrating the testis, which suggests that these cells have the potential to be infected by HIV. Indeed, in 9 of the 23 cases in which immunocytochemistry was used to test for HIV-1 protein, HIV-infected cells of the lymphocytic/monocytic type were found in the seminiferous tubules and interstitium of the testis. Such cells were also found in the semen (257). Using in situ PCR, several other studies have detected HIV-1 DNA within testicular germ cells. (i) The infection of spermatogonia, spermatocytes, and a small number of spermatids in 11 of 12 men with AIDS was described by Nuovo et al. (240). (ii) Another study with seropositive asymptomatic subjects reported the presence of HIV-1 DNA in the nuclei of spermatogonia and other germ cells at all stages of differentiation, suggesting clonal infection (228). In these subjects, the presence of provirus did not cause cell damage and was associated with normal spermatogenesis, whereas in AIDS patients, spermatogenesis was arrested and there were few infected spermatogonia and spermatocytes. (iii) HIV-1 DNA was recently found in 25 to 33% of the residual germ cells in the testes of AIDS patients, with testicular changes including all variants from decreased spermatogenesis to the Sertoli-cell-only pattern (303). In contrast, HIV-1 DNA was not detected in the testes of any of three preadolescent boys who acquired HIV in utero (303). The way in which HIV enters germ cells is unknown. The CD4 receptor has still not been unequivocally detected on the germ cell surface, nor have the chemokine coreceptors triggering HIV entry in other cell types (e.g., monocytes/macrophages, lymphocytes and microglial cells). The virus may also enter germ cells via an alternative galactoglycerolipid receptor, as suggested (52). Tissue-specific HIV-1 quasispecies have been identified in the testis, indicating that there may be a tissue tropism for this organ (341). This is a very important issue for HIV treatment because triple therapy seems to be unable to eradicate the virus completely. Thus, the testis could act as a viral reservoir isolated by the blood-testis barrier which cannot be reached by drugs (71) (see the section on HIV-1 in Human semen above). Oncogenic viruses. Orchitis has been described as a symptom of infectious mononucleosis caused by EBV, indicating that this virus may have a direct effect on the testis (268). EBV DNA was detected in the testes of three individuals with no EBV-related disease, showing that the testis may be a target organ (67). Rajpert De Meyts et al. (262) found that EBV was not directly involved in the testicular germ cell tumors of 20 patients but thought that germ cell proliferation might be stimulated by testicular EBV-transformed lymphocytes. However, more recently, Shimakage et al. (305) detected EBV RNA and EBV-related proteins in all of the 27 seminomas they tested but in none of 25 nonmalignant testes, providing further support for the hypothesis that EBV is associated with testicular tumors. In addition, transgenic mice expressing papillomavirus genes develop germ cell tumors resembling human seminoma (171). Recently, 39 patients with testicular germ cell tumors were screened for EBV, CMV, and parvovirus B19 DNA by PCR. Neither EBV nor CMV DNA was detected in the testis. In contrast, parvovirus B19 DNA sequences were found in the testicular tissues of 85% of patients with testicular cancers but in none of the normal testicular tissue samples (129). However, the cellular targets of parvovirus B19 within normal and cancerous testes are unknown and the possible role (direct or indirect) of this virus in the development of testicular germ cell tumors is unclear. Other viruses. HSV-2 was detected in the testes of 4 of 10 corpses at autopsy, suggesting that this organ acts as a reservoir for transmission of the virus (95). Csata and Kulcsar (80) studied the relationship between male infertility and the presence in the semen or testis of HSV-1 or adenovirus. In 40% of patients with infertility either HSV-1 or adenovirus was present within the testis. Testicular cells were also infected in vitro by either HSV-1 or adenovirus, and these two viruses were found to penetrate and replicate (80). Several viruses are suspected of inducing orchitis, based on